Serum Tau,Neurofilament Light Chain,Glial Fibrillary Acidic Protein,and Ubiquitin Carboxyl-Terminal Hydrolase L1 Are Associated with the Chronic Deterioration of Neurobehavioral Symptoms after Traumatic Brain Injury

Abstract

The purpose of this study was to examine the association of serum tau, neurofilament light chain (NFL), glial fibrillary acidic protein (GFAP), and ubiquitin carboxy-terminal hydrolase L1 (UCHL-1) concentrations evaluated within the first 12 months after a military-related TBI, with longitudinal changes in neurobehavioral functioning extending two or more years post-injury. Participants were 84 United States service members and veterans (SMVs) prospectively enrolled in the Defense and Veterans Brain Injury Center of Excellence/Traumatic Brain Injury Center 15-Year Longitudinal TBI Study, separated into three discreet groups: (a) uncomplicated mild TBI (MTBI; n = 28), (b) complicated mild, moderate, severe, and penetrating TBI combined (STBI; n = 29], and (c) non-injured controls (NIC, n = 27). Participants completed a battery of self-report neurobehavioral symptom measures (e.g., depression, post-traumatic stress disorder [PTSD], post-concussion, anxiety, somatic, cognitive, and neurological symptoms) within 12 months of injury (baseline), and then again at two or more years post-injury (follow-up). At baseline, participants also completed a blood draw to determine serum concentrations of tau, NFL, GFAP, and UCHL-1 using an ultra-sensitivity assay method. In the MTBI and STBI groups (using hierarchical regression analyses), (1) baseline tau concentrations predicted the deterioration of neurobehavioral symptoms from baseline to follow-up on measures of anxiety, PTSD, depression, post-concussion, somatic, and neurological symptoms (accounting for 10–28% of the variance); (2) NFL predicted the deterioration of depression, post-concussion, somatic, cognitive, and neurological symptoms (10–32% variance); (3) GFAP predicted the deterioration of post-concussion, PTSD, depression, anxiety, somatic, neurological, and cognitive symptoms (11–43% variance); and (4) UCHL-1 predicted the deterioration of anxiety, somatic, and neurological symptoms (10–16% variance). In the NIC group, no meaningful associations were found between baseline biomarker concentrations and the deterioration of neurobehavioral symptoms on the majority of measures. This study reports that elevated tau, NFL, GFAP, and UCHL-1 concentrations within the first 12 months of injury are associated with the deterioration of neurobehavioral symptoms that extends to the chronic phase of recovery after a TBI. These findings suggest that a blood-based panel including these biomarkers could be a useful prognostic tool to identifying those individuals at risk of poor future outcome after TBI.

Introduction

The role of protein biomarkers as diagnostic and prognostic indicators after traumatic brain injury (TBI) has received considerable attention. Evidence accumulated over the past decades indicates that TBI triggers cellular and molecular processes that continue to evolve after the acute period of recovery. While these mechanisms result in tissue repair and regeneration, the reparative process is often incomplete and can also be maladaptive if it is not sufficiently regulated.¹

To date, the majority of the literature in this area has focused on demonstrating that a variety of protein biomarkers (including serum tau, neurofilament light chain [NFL], glial fibrillary acidic protein [GFAP], and ubiquitin carboxyl-terminal hydrolase L1 [UCHL-1]) have clinical utility to (a) differentiate individuals with TBI versus those without, (b) distinguish TBI severity groups, (c) identify positive versus negative neuroimaging scans after mild TBI, or (d) are associated with mortality rates and/or disability severity ratings after moderate to severe TBI (see ^1

–5 for reviews).

Recent research has also expanded the potential diagnostic and prognostic value of protein biomarkers by establishing an association between serum tau, NFL, GFAP, and/or UCHL-1 with more subtle neurobehavioral outcomes commonly associated with the sequelae of TBI such as post-concussion symptoms, post-traumatic stress disorder (PTSD), depression, sleep problems, and cognitive dysfunction (e.g., ^{6

–16}), although not all articles have found consistent results (e.g.,^7,9,17).

The vast majority of the studies to date are cross-sectional in design and/or do not provide information regarding the prognostic value of biomarkers to predict long-term future neurobehavioral outcome after TBI. Nonetheless, there are some notable exceptions. There is an important body of research that has demonstrated that serum tau, NFL, GFAP, and/or UCHL-1 evaluated within the first few hours, days, or weeks after moderate or severe TBI (and mild TBI in some studies) is associated with worse disability severity ratings as measured by the Glasgow Outcome Scale (GOS) or GOS-Extended (GOS-E) at three, six, or 12 months post-injury (e.g., ^18
–20). These studies, however, do not provide information regarding more subtle long-term neurobehavioral outcomes commonly associated with the sequelae of TBI (e.g., depression, sleep).

In addition, there is an important body of research in athletes that has found that serum tau, GFAP, and NFL evaluated within the first few hours/days after a sport-related concussion is associated with prolonged return to play within the first 10–14 days post-injury.^21,22 These studies, however, focus on acute activity of these proteins, and subacute clinical outcomes, which tend not to be applicable to a large portion of patients who have sustained a TBI who are assessed later in the recovery trajectory.

To our knowledge, there is only one longitudinal study that has examined the association between serum biomarkers to predict future neurobehavioral outcomes in the chronic phase of recovery after TBI. Hossain and colleagues¹⁸ examined NFL and GFAP concentrations evaluated within 24 h after uncomplicated and complicated mild TBI. While these authors found modest predictive utility of baseline NFL and GFAP concentration levels to predict GOS-E scores at follow-up 6–12 months post-injury (e.g., odds ratio [OR] = 1.009), they did not find any relationship between NFL and GFAP concentrations with post-concussion symptoms 6–12 months post-injury.

The purpose of this study was to examine the association of serum tau, NFL, GFAP, and UCHL-1 concentrations within the first 12 months after TBI with longitudinal changes in neurobehavioral functioning two or more years post-injury. This study will expand on previous longitudinal research in this area by (a) evaluating the prognostic utility of serum biomarkers evaluated within the first 12 months after injury (vs. the first few days/weeks post-injury), (b) examine outcome from TBI two or more years post-injury (vs. within the first 6–12 months post-injury), and (c) include a broad range of neurobehavioral outcomes commonly reported after TBI (e.g., depression, PTSD, post-concussion, cognitive, somatic symptoms).

It is hypothesized that higher serum biomarker concentrations within the first 12 months after TBI will be associated with the deterioration of neurobehavioral functioning two or more years post-injury.

Methods

Participants

Participants were 84 United States service members and veterans (SMVs) prospectively enrolled in the Defense and Veterans Brain Injury Center [DVBIC]/Traumatic Brain Injury Center of Excellence [TBICoE] 15-Year Longitudinal TBI Study. Participants were recruited into three groups (uncomplicated mild TBI [MTBI; n = 28], complicated mild, moderate, severe, and penetrating TBI combined [STBI; n = 29], and non-injured controls [NIC, n = 27]) from inpatient and outpatient wards at the Walter Reed National Military Medical Center (WRNMMC), Bethesda, MD (82.1%), as well as via community-based recruitment initiatives (17.9%).

Criteria for inclusion were as follows: MTBI group: (i) post-traumatic amnesia (PTA) ≤24 h, loss of consciousness (LOC) ≤30 min, and/or alteration of consciousness (AOC) present, and (ii) no trauma-related intracranial abnormality (ICA); STBI group: met criteria for one of the following: (a) complicated MTBI (n = 11): (i) Glasgow Coma Scale (GCS) score = 13–15, PTA <24 h, LOC <30 min, and/or AOC present, and (ii) trauma-related ICA on computed tomography (CT) or magnetic resonance imaging (MRI); (b) moderate TBI (n = 8): LOC 30 min–24 h, PTA 1–7 days, and/or lowest reliable GCS score (e.g., not intubated/sedated/intoxicated) >30 min post-injury of 9–12; (c) severe TBI (n = 7): LOC >24 h, PTA >7 days, and/or lowest reliable GCS score >30 min post-injury of <9; (d) penetrating TBI (n = 3): a breach of the cranial vault and/or dura mater by an external object (e.g., bullet, shrapnel) and/or skull fragment (i.e., depressed skull fracture). NIC group: no history of an orthopedic/soft-tissue injury or TBI.

Participants were excluded if they were not proficient in reading or conversing in English; had a history of an unrelated neurological (e.g., meningioma) or psychiatric condition (e.g., bipolar disorder); or were less than 18 years of age.

Participants were selected from a larger sample of patients who had been enrolled in the DVBIC-TBICoE 15-Year Longitudinal TBI study if they (a) had completed two evaluations (i.e., baseline and follow-up), (b) had completed a baseline evaluation within 12 months post-injury (TBI only), (c) had completed a follow-up evaluation two or more years post-injury injury (TBI only), (d) had completed the target self-report symptom measures at baseline and follow-up, (e) had scored below the recommended cutoff (at baseline and follow-up) on a measure designed to evaluate symptom exaggeration) see Measures section], and (f) had at least one usable blood sample at baseline (see Measures section).

The protocol under which these data were collected was approved by the Institutional Review Board of WRNMMC, Bethesda, MD. This study was completed in accordance with the guidelines of the Declaration of Helsinki. All participants provided informed consent.

TBI diagnosis and classification

Diagnosis and classification of TBI was based on a medical record review and a comprehensive lifetime TBI history interview. The TBI history interview consisted of the Ohio State University TBI identification method²³ and an extended semi-structured clinical interview designed to (a) extract more detailed information to estimate the presence/duration of LOC, PTA, AOC, and retrograde amnesia, and (b) gather military-specific information regarding injury circumstances (e.g., type of blast, protection worn, etc.). Final determination and classification of TBI severity was undertaken by consensus, giving consideration to all information, during case conferencing with the interviewer and a PhD-level clinician/scientist trained in neuropsychology and TBI diagnostic interviewing (RTL or SML).

Measures and procedure

Laboratory analyses

Non-fasting blood samples were collected with plastic lithium heparin tubes, processed within an hour, and stored at -80°C. Batch assays were conducted after all samples had been collected. Simoa™ (Quanterix, Lexington, MA), a high-definition-1 analyzer, was used to measure serum concentrations. Blinded laboratory scientists ran the 4-Plex assay randomized over plates and in duplicate.

Blood samples were not used if the coefficients of variation (CV) were more than 20% and the levels were above the lower limit of quantification (mean CVs: tau = 9.2%, NFL = 6.3%, GFAP = 3.4%, UCHL-1 = 17.5%). The lower limit of quantification for the assay is as follows: tau = 0.053 pg/mL, NFL = 0.241 pg/mL, GFAP = 0.467 pg/mL, and UCHL-1 = 5.45 pg/mL. Participants were required to have at least one useable sample based on these criteria to be included in the study. In the selected sample, 100% of participants had usable NFL and GFAP samples, 79.8% had a usable tau sample, and 52.4% had a usable UCHL-1 sample.

Self-reported neurobehavioral symptoms

The Minnesota Multiphasic Personality Inventory-2nd Edition-Restructured Format [MMPI-2-RF]²⁴ is a 338-item measure designed to evaluate psychological symptomatology. Responses of the MMPI-2-RF are used to generate norm-based T-scores for eight Validity scales, three Higher Order scales, nine Clinical scales, and 25 Specific Problem scales.

For the purpose of this study, only nine scales were selected for examination based on symptoms that are commonly reported after TBI: (i) Demoralization (RCd), (ii) Somatic Complaints (RC1), (iii) Low Positive Emotions (RC2), (iv) Malaise (MLS), (v) Head Pain Complaints (HPC), (vi) Neurological Complaints (NUC), (vii) Cognitive Complaints (COG), (viii) Anxiety (AXY), and (ix) Anger Proneness (ANP). In addition, the validity scales were used for the purposes of evaluating symptom validity.

Based on recommended cutoff scores, participants were not included if they were considered to have exaggerated symptoms on the MMPI-2-RF (i.e., F-r ≥100T or Fp-r ≥90T or Fs ≥100T or FBS-r ≥100T or RBS ≥100T) or whose scores were not considered interpretable (i.e., Cannot Say scores >14, or VRIN-r/TRIN-r scores >79T).

The Neurobehavioral Symptom Inventory [NSI]²⁵ is a 22-item measure designed to evaluate self-reported post-concussion symptoms (e.g., headache, balance, nausea, etc.) rated on a 5-point scale. A total score can be obtained by summing the ratings for the 22 items, in addition to the calculation of four cluster scores as outlined by Vanderploeg and colleagues²⁶: i.e., Vestibular, Somatosensory, Cognitive, and Affective clusters.

The PTSD Checklist-Civilian version (PCLC)²⁷ is a 17-item measure patterned specifically after the DSM-IV-TR²⁸ symptom criteria for PTSD. A total score can be obtained by summing the ratings for the 17 items, in addition to the calculation of three cluster scores that correspond to the DSM-IV-TR symptom criteria: Criterion B (Re-Experiencing cluster), Criterion C (Avoidance cluster), and Criterion D (Hyperarousal cluster).

Calculation of symptom change scores

For the purposes of evaluating change in symptom expression over time, “symptom change” scores were calculated (continuous variable) by subtracting scores at follow-up from scores at baseline for each participant. Negative symptom change scores indicate a deterioration of symptoms from baseline to follow-up. Positive symptom change scores indicate an improvement of symptoms from baseline to follow-up.

For some analyses, symptom changes scores were classified as representing the presence or absence of a deterioration of symptoms from baseline to follow-up. The presence of symptom deterioration was defined as follows: (a) MMPI-2-RF scales = symptom change score ≥1 SD (i.e., ≥ -10 T-score points); and (b) NSI/PCLC measures = symptom change scores that exceeded Reliable Change Index (RCI) cutoff scores (i.e., 90% confidence Interval) developed by Belanger and colleagues.²⁹

The cutoff scores applied for the NSI/PCLC measures was as follows: NSI total ≥ -8, Vestibular ≥ -2, Somatosensory ≥ -4, Cognitive ≥ -3, Affective ≥ -4; PCLC total ≥ -7, Re-experiencing ≥ -2, Avoidance ≥ -3, and Hyperarousal ≥ -2. Note that Belanger and colleagues²⁹ did not provide RCI cutoff scores for the three PCLC clusters. As such, the cutoff scores for the three PCLC clusters were prorated from the PCLC total score based on the number of items in each cluster.

Statistical analyses

Descriptive statistics and group comparisons were undertaken using analyses of variance (ANOVAs, i.e., demographic variables, injury variables, neurobehavioral measures) or Kruskal-Wallis H tests (i.e., biomarkers) for continuous variables, and chi-square analysis for categorical variables (i.e., demographic measures, prevalence of symptom deterioration). Pairwise comparisons were undertaken using Fisher least significant difference (i.e., demographics, self-report measures) or Mann-Whitney U tests (i.e., biomarkers) for continuous variables, and chi-square analysis for categorical variables. Cohen effect sizes d or H were calculated where appropriate.

Because of the small sample size (i.e., low statistical power), when p values failed to reach threshold for statistical significance (i.e., p < 0.05), meaningful group differences were interpreted using Cohen effect sizes, as outlined by Corrigan and coworkers,³⁰ who recommended that an effect size of >0.30 was considered “very important.” Spearman rho correlation analysis (continuous variables) and Mann-Whitney U tests (categorical variables) were used to examine the relationship between the four biomarkers and select demographic and injury characteristics (i.e., age, education, gender, ethnicity, number of deployments, time since injury, number of lifetime TBIs). Any demographic or injury variable that was meaningfully associated with a biomarker (i.e., r_s ≥ 0.30 for continuous variables, or p < 0.05 for categorical variables) was used as a covariate in subsequent analyses.

A series of hierarchical regression analyses were undertaken, including relevant covariates, to examine whether each of the four biomarkers at baseline could predict symptom change scores on each of the neurobehavioral measures separately. Because of small sample sizes, a clinically meaningful relationship between a biomarker and symptom change score in the regression analyses was defined as either (a) p < 0.05 or (b) R²/R²Δ > 0.10 (i.e., percent variance accounted for >10%). The criterion for R²/R²Δ was chosen because statistically significant results using regression analyses are routinely found in larger samples when R²/R²Δ < 0.10 (e.g., ^11,2).

Results

Descriptive statistics and group comparisons across baseline demographic and injury variables are presented in Table 1. There were significant main effects for gender and ethnicity, but not for age, education, number of deployments, time since injury, or number of lifetime TBIs. Pairwise comparisons revealed that there was a significantly higher proportion of participants in the MTBI and STBI groups who were male and white, compared with the NIC group.

Table 1.

Descriptive Statistics and Group Comparisons for Baseline Demographic and Injury Characteristics

	1. NIC		2. MTBI		3. STBI		ANOVA p	Pairwise comparisons (p < .05)
	M	SD	M	SD	M	SD	ANOVA p	Pairwise comparisons (p < .05)
Age	34.6	10.1	32.2	9.1	32.7	10.3	.641	-
Education	15.1	2.0	14.7	2.1	14.4	2.0	.477	-
Number of deployments	2.3	4.8	2.4	3.3	2.2	1.7	.971	-
Time since injury (baseline)	-	-	6.9	4.4	8.3	3.9	.225	-
Time since injury (follow-up)	-	-	42.6	11.3	40.2	10.1	.395	-
Number of lifetime TBIs	-	-	1.3	0.5	1.1	0.5	.383	-

	N	%	n	%	n	%	K-W p	Pairwise comparisonsM-W (p < .05)
Gender (male)	19	70.4	27	96.4	27	93.1	.007	1 < 2 & 3
Ethnicity/race (white)	15	55.6	21	75.0	24	82.8	.065	1 < 2 & 3

NIC, non-injured control; MTBI. uncomplicated mild TBI; STBI, uncomplicated mild, moderate, severe, and penetrating TBI combined; ANOVA, analysis of variance; SD, standard deviation; TBI, traumatic brain injury; K-W, Kruskal-Wallis H test; M-W, Mann-Whitney U test.

N = 84 (NIC = 27, MTBI = 28, STBI = 29).

Descriptive statistics and group comparisons across the four biomarker concentrations and self-report symptom measures at baseline are presented in Table 2. For the biomarkers, there were significant main effects for NFL, GFAP, and UCHL-1, but not for tau. Pairwise comparisons revealed higher concentrations of GFAP and UCHL-1 in the NIC group compared with the MTBI and STBI groups. In addition, the STBI group had higher concentrations of tau compared with the NIC group.

Table 2.

Descriptive Statistics and Group Comparison of Baseline Biomarker Concentrations and Neurobehavioral Measures

	1. NIC		2. MTBI		3. STBI		K-W (p)	Pairwise comparisons (M-W)			Summary (p < .05)
	1. NIC		2. MTBI		3. STBI			1 v 2	1 v 3	2 v 3
	Med	IQR	Med	IQR	Med	IQR		p	p	p
Tau¹	0.6	0.4-0.8	0.5	0.3-0.8	0.4	0.3-0.8	.538	.378	.298	.818	-
NFL	6.5	5.2-9.5	7.3	5.0-12.9	11.6	6.9-18.2	.026	.511	.007	.082	1 < 3
GFAP	87.0	64.7-112.5	51.6	41.2-82.5	69.4	48.1-82.0	<.001	<.001	.024	.074	1 > 2 & 3
UCHL1	14.5	10.6-33.2	5.1	4.6-13.6	5.2	4.1-9.0	.005	.026	<.001	.483	1 > 2 & 3

							ANOVA(p)	Pairwise comparisons						Summary(p < .05 or d > .30)
Measures	M	SD	M	SD	M	SD	ANOVA(p)	d	p	d	p	d	p	Summary(p < .05 or d > .30)
MMPI-2-RF
Demoralization	48.8	8.9	47.9	11.2	49.8	10.7	.773	.09	.728	.10	.703	.18	.500	-
Somatic complaints	51.0	10.4	57.0	8.8	57.9	8.8	.014	.63	.024	.73	.009	.11	.691	1 < 2 & 3
Low positive emotions	48.9	10.8	49.1	10.6	51.4	13.0	.666	.03	.920	.21	.434	.19	.481	-
Malaise	49.9	7.6	57.3	11.3	58.4	12.0	.006	.78	.006	.87	.003	.10	.707	1 < 2 & 3
Head pain complaints	50.1	10.2	53.9	11.6	56.3	10.6	.104	.35	.197	.59	.030	.21	.428	1 < 2 & 3
Neurological complaints	51.9	12.3	61.6	9.3	60.6	10.7	.002	.90	.002	.76	.006	.10	.713	1 < 2 & 3
Cognitive complaints	53.5	13.3	61.3	14.8	58.6	13.9	.116	.56	.044	.38	.164	.19	.480	1 < 2 & 3
Anxiety	47.7	7.4	48.9	9.5	52.1	12.6	.250	.14	.618	.43	.124	.29	.284	1 < 3
Anger proneness	49.4	9.0	50.0	10.5	52.0	10.3	.578	.06	.812	.28	.308	.20	.463	-
NSI
Total	6.7	8.2	14.0	12.9	16.3	14.2	.012	.69	.015	.85	.004	.17	.532	1 < 2 & 3
Vestibular	0.3	0.7	1.0	1.4	1.4	1.8	.013	.73	.013	.87	.004	.20	.457	1 < 2 & 3
Somato-Sensory	1.6	2.4	3.4	3.8	4.0	4.3	.041	.57	.045	.72	.013	.16	.554	1 < 2 & 3
Cognitive	1.6	2.2	3.9	3.9	3.4	3.5	.032	.76	.010	.61	.034	.15	.568	1 < 2 & 3
Affective	2.6	3.2	4.8	4.4	6.8	6.0	.006	.57	.040	.90	.002	.38	.167	1 < 2 < 3
PCLC
Total	21.0	5.1	24.3	8.6	27.8	12.2	.026	.47	.095	.78	.010	.34	.214	1 < 2 < 3
Re-experiencing	5.5	1.6	6.4	2.7	7.7	3.8	.024	.40	.155	.79	.009	.39	.154	1 < 2 < 3
Avoidance	8.4	2.1	9.0	3.0	10.6	4.9	.066	.25	.370	.62	.037	.40	.155	1 & 2 < 3
Hyperarousal	7.1	2.2	8.9	4.0	9.5	4.2	.040	.56	.050	.75	.010	.17	.534	1 < 2 & 3

NIC, non-injured control; MTBI, uncomplicated mild TBI; STBI, uncomplicated mild, moderate, severe, and penetrating TBI combined; K-W, Kruskal-Wallis H test; M-W, Mann-Whitney U test; Med, median; IQR, interquartile range; NFL, neurofilament light; GFAP, glial fibrillary acidic protein; UCHL-1, ubiquitin carboxyl-terminal hydrolase L1; SD, standard deviation; ANOVA, analysis of variance; NSI, Neurobehavioral Symptom Inventory; PCLC, PTSD Checklist; TBI, traumatic brain injury.

N = 84 (NIC = 27, MTBI = 28, STBI = 29).

Usable samples by group: (a) Tau = 67 (NIC = 21, MTBI = 23, STBI = 23); (b) NFL = 84 (NIC = 27, MTBI = 28, STBI = 29); (c) GFAP = 84 (NIC = 27, MTBI = 28, STB I = 29); (d) UCHL-1 = 44 (NIC = 14, MTBI = 17, STBI = 13).

For the self-report measures, there were significant main effects for 11 of the 18 measures. Pairwise comparisons revealed that the MTBI and STBI groups had consistently higher scores on RC1, MLS, HPC, NUC, and COG scales, as well as NSI total, NSI Vestibular, NSI Somatosensory, NSI Cognitive, and PCLC Hyperarousal. In addition, the STBI group had higher scores on NSI Affective, PCLC total, and PCLC Re-experiencing compared with the MTBI group, with the MTBI group having higher scores on these same measures compared with the NIC group.

Descriptive statistics and group comparisons for the prevalence of deteriorated symptoms from baseline to follow-up are presented in Table 3. There were statistically significant main effects for two of the 18 measures (NSI Vestibular and NSI Cognitive) and for a handful of pairwise comparisons (i.e., MTBI > NIC [NSI Vestibular and NSI Somatosensory]; STBI > NIC [NUC, NSI Vestibular, NSI Somatosensory, and Cognitive]; and STBI > MTBI [RC2, NSI Affective]). Because of the small sample size (i.e., low statistical power), however, many meaningful group differences were not identified using p values as a criterion.

Table 3.

Descriptive Statistics and Group Comparisons for the Percent of Deteriorated Symptoms from Baseline to Follow-up

	1. NIC	2. MTBI	3. STBI	Exact Tests (p)	Pairwise comparisons						Summary (H ≥ .30)
	1. NIC	2. MTBI	3. STBI		1 v 2		1 v 3		2 v 3
Measures	%	%	%		%diff	H¹	% diff	H	% diff	H
MMPI-2-RF
Demoralization SxΔ cat²	18.5	21.4	34.5	.375	2.9	.08	16.0^	.37	13.1	.29	1 < 3
Somatic complaints SxΔ cat	14.8	17.9	24.1	.736	3.1	.08	9.3	.23	6.2	.15	–
Low positive emotions SxΔ cat	25.9	10.7	34.5	.099	15.2^	.41	8.6	.19	23.8^*	.60	2 < 1 & 3
Malaise SxΔ cat	14.8	21.4	34.5	.236	6.6	.17	19.7^	.46	13.1	.29	1 < 3
Head pain complaints SxΔ cat	25.9	32.1	31.0	.872	6.2	.13	5.1	.11	-1.1	.02	–
Neurological complaints SxΔ cat	7.4	17.9	27.6	.148	10.5^	.32	20.2^*	.55	9.7	.23	1 < 2 & 3
Cognitive complaints SxΔ cat	22.2	32.1	37.9	.464	9.9	.23	15.7^	.35	5.8	.13	1 < 3
Anxiety SxΔ cat	25.9	25.0	37.9	.525	-0.9	.02	12	.26	12.9	.28	–
Anger proneness SxΔ cat	11.1	21.4	27.6	.323	10.3	.29	16.5^	.43	6.2	.14	1 < 3
NSI
Total SxΔ cat	14.8	21.4	32.1	.333	6.6	.17	17.3^	.41	10.7	.23	1 < 3
Vestibular SxΔ cat	0.0	21.4	21.4	.019	21.4^*	.97	21.4^*	.97	0.0	.00	1 < 2 & 3
Somato-Sensory SxΔ cat	0.0	14.3	17.9	.069	14.3^*	.78	17.9^*	.88	3.6	.10	1 < 2 & 3
Cognitive SxΔ cat	7.4	10.7	35.7	.017	3.3	.11	28.3^*	.72	25.0^*	.62	1 & 2 < 3
Affective SxΔ cat	22.2	21.4	28.6	.852	-0.8	.01	6.4	.15	7.2	.16	–
PCLC
Total SxΔ cat	18.5	14.3	32.1	.299	-4.2	.11	13.6^	.31	17.8^	.42	1 & 2 < 3
Re-experiencing SxΔ cat	14.8	25.0	32.1	.363	10.2	.25	17.3^	.41	7.1	.16	1 < 3
Avoidance SxΔ cat	18.5	21.4	39.3	.202	2.9	.08	20.8^	.47	17.9^	.39	1 & 2 < 3
Hyperarousal SxΔ cat	25.9	39.3	50.0	.192	13.4	.29	24.1^	.50	10.7	.21	1 < 3

NIC, non-injured control; MTBI, uncomplicated mild TBI; STBI, complicated mild, moderate, severe, and penetrating TBI combined; MMPI-2-RF, Minnesota Multiphasic Personality Inventory-2nd Edition-Restructured Format; NSI, Neurobehavioral Symptom Inventory; PCLC, PTSD Checklist.

N = 84 (NIC = 27, MTBI = 28, STBI = 29).

^meaningful difference, H ≥ .30; ^* p < .0.;

Cohen Effect Size H: small = 0.2, medium = 0.5, large = 0.8.

The presence of symptom deterioration was defined as follows: (a) MMPI-2-RF scales, symptom change score of ≥1 SD (i.e., ≥ -10 T-score points); and (b) NSI/PCLC measures, symptom change scores that exceeded Reliable Change Index (RCI) cutoff scores (i.e., 90% confidence interval) developed by Belanger et al., 2016: NSI total ≥ -8, Vestibular ≥ -2, Somatosensory ≥ -4, Cognitive ≥ -3, Affective ≥ -4 ; PCLC total ≥ -7, Re-experiencing ≥ -2, Avoidance ≥ -3, and Hyperarousal ≥ -2.

Examination of Cohen effect sizes (i.e., ≥0.30) revealed that a higher proportion of the STBI group experienced a deterioration of symptoms from baseline to follow-up on 13 of the 18 measures compared with the NIC group (e.g., RCd, MLS, COG, ANP; NSI total, PCLC total; H = 0.31 to H = 0.97), and on four of the 18 measures compared with the MTBI group (RC2, NSI Cognitive, PCLC total, PCLC Avoidance; H = 0.39 to H = 0.62). In addition, a higher proportion of the MTBI group experienced a deterioration of symptoms from baseline to follow-up on four of the 18 measures compared with the NIC group (RC2, NUC; NSI Vestibular and NSI Somatosensory; H = 0.32 to H = 0.97).

Spearman rho correlations and Mann-Whitney U tests examining the association between select demographic and injury variables with baseline biomarkers in the entire sample are presented in Table 4. There were meaningful correlations (r_s ≥ 0.30) between NFL and time since injury (r_s = -0.39), and GFAP with age (r_s = 0.30) and education (r_s = 0.32). For UCHL-1, there was a meaningful correlation with age (r_s = 0.31), as well as a significant group difference for ethnicity (p = 0.031). For tau, there were no meaningful correlations or group differences for all variables.

Table 4.

Spearman Rho Correlation Coefficients and Mann-Whitney U Tests Comparing Select Demographic and Injury Variables with Baseline Biomarkers in the Entire Sample

	Tau	NFL	GFAP	UCHL-1
	r_s	r_s	r_s	r_s
Age	-.06	.05	.30^**	.31^**
Education	.10	.18	.32^**	.12
Number of deployments	-.09	.10	-.14	-.17
Time since injury¹	.09	-.39^**	.01	-.05
Number of TBI²	.02	.01	.03	-.10

	p	p	p	p
Ethnicity (White/Non-White)	.210	.243	.231	.031^*
Gender (Male/Female)	.109	.235	.378	.489

NFL, neurofilament light; GFAP, glial fibrillary acidic protein; UCHL-1, ubiquitin carboxyl-terminal hydrolase l1; TBI, traumatic brain injury.

N, 84 (NIC, 27, MTBI, 28, STBI, 29).

p < .05, ^** p < .001.

A summary of the hierarchical regression analyses (adjusted for relevant covariates) using the four biomarkers at baseline to predict symptom change scores on each measure (i.e., baseline to follow-up) by group is presented in Table 5. Overall, all four baseline biomarkers were predictive of symptom change scores on multiple measures in both the MTBI and STBI groups. Baseline tau concentrations were predictive of symptom change scores on measures of (a) somatic (RC1, HPC) and neurological (NUC, NSI Vestibular) complaints, anxiety (AXY), and PTSD (PCLC total, PCLC Re-experiencing, Avoidance) in the MTBI group; and (b) depression (RC2), somatic (RC1, MLS, NSI Somatosensory) and neurological complaints (NSI Vestibular), and post-concussion symptoms (NSI total) in the STBI group (accounting for 10.0- 27.7% of the variance).

Table 5.

Summary of Regression Analyses for Baseline Tau and NFL Concentrations To Predict Neurobehavioral Difference Scores from Baseline to Follow-Up by Group

	Tau			NFL			GFAP			UCHL1
	NIC	MTBI	STBI	NIC	MTBI	STBI	NIC	MTBI	STBI	MSTBI
Measures	R²	R²	R²	R²	R²Δ	R²Δ	R²Δ	R²Δ	R²Δ	R²Δ
MMPI-2-RF
Demoralization SxΔ	.099	.071	.052	.034	.010^	.318^**	.001	.045	.206^**	.079
Somatic complaints SxΔ	.001	.101^	.123^	.004	.005	.077	.049	.054	.063	.069
Low positive emotions SxΔ	.016	.067	.224^*	.001	.006	.038	.027	.229^**	.008	.094
Malaise SxΔ	.036	.023	.203^*	.072	.086	.055	.068	.346^***	.003	.124^
Head Pain complaints SxΔ	.027	.239^*	.062	.005	.009	.028	.027	.076	.002	.066
Neurological complaints SxΔ	.001	.152^	.066	.049	.007	.103	.065	.015	.106^	.164^*
Cognitive complaints SxΔ	.082	.012	.006	.003	.037	.196^*	.016	.040	.110^	.010
Anxiety SxΔ	.000	.256^*	.001	.139^	.009	.041	.014	.011	.076	.103^
Anger proneness SxΔ	.272^*	.008	.002	.003	.001	.005	.057	.016	.021	.002
NSI
Total SxΔ	.020	.044	.154^	.038	.001	.140^*	.014	.027	.290^**	.001
Vestibular SxΔ	.001	.146^	.277^*	.007	.010	.236^**	.002	.012	.262^**	.066
Somato-Sensory SxΔ	.001	.072	.132^	.034	.006	.150^*	.001	.006	.430^***	.022
Cognitive SxΔ	.002	.027	.060	.014	.001	.131^	.062	.050	.198^*	.003
Affective SxΔ	.112	.004	.053	.034	.067	.061	.001	.045	.135^*	.014
PCLC
Total SxΔ	.011	.131^	.017	.013	.018	.042	.024	.014	.120^	.007
Re-experiencing SxΔ	.001	.232^*	.002	.013	.049	.000	.025	.001	.102^	.016
Avoidance SxΔ	.013	.100^	.009	.007	.000	.041	.023	.024	.063	.014
Hyperarousal SxΔ	.061	.017	.035	.017	.012	.099	.008	.019	.146^*	.001

NFL, neurofilament light; GFAP, glial fibrillary acidic protein; UCHL-1, ubiquitin carboxyl-terminal hydrolase L1; NIC, non-injured control; MTBI, uncomplicated mild TBI; STBI, uncomplicated mild, moderate, severe, and penetrating TBI combined; MMPI-2-RF, Minnesota Multiphasic Personality Inventory-2nd Edition-Restructured Format; SxΔ = Symptom Change score calculated between Baseline and Follow-up evaluations; NSI, Neurobehavioral Symptom Inventory; PCLC, PTSD Checklist.

N = 84 (NIC = 27, MTBI = 28, STBI = 2 9).

^Meaningful relationship-percent variance ≥10%.

p < .05, ^** p < .01, ^*** p < .001.

Usable samples by group: (a) Tau = 67 (NIC = 21, MTBI = 23, STBI = 23); (b) NFL = 84 (NIC = 27, MTBI = 28, STBI = 29); (c) GFAP = 84 (NIC = 27, MTBI = 28, STBI = 29); (d) UCHL1 = 44 (MTBI+STBI = 30; NIC = 14, MTBI = 17, STBI = 13 [note that data for the NIC, MTBI, and STBI were not included because of sample sizes that were considered too small for reliable analyses]).

Covariates included: Tau = none; NFL = time since injury (TBI group only); GFAP = age and education; UCHL1 = age and ethnicity.

Baseline NFL concentrations were predictive of symptom change scores on measures of (a) depression (RCd) in the MTBI group; and (b) depression (RCd), somatic (NSI Somatosensory), cognitive (COG, NSI Cognitive), and neurological complaints (NSI Vestibular]), and post-concussion symptoms (NSI total) in the STBI group (accounting for 10.0–31.8% of the variance).

Baseline GFAP concentrations were predictive of symptom change scores on measures of (a) depression (RC2) and somatic complaints (MLS, NSI Somatosensory) in the MTBI group; (b) depression and anxiety (RCd, NSI Affective), somatic (NSI Somatosensory), neurological (NUC, NSI Vestibular) and cognitive (COG, NSI Cognitive) complaints, post-concussion symptoms (NSI total), and PTSD (PCLC total, Re-experiencing, Avoidance) in the STBI group (accounting for 10.6–43.0% of the variance).

For UCHL-1, the sample sizes in the three groups with useable UCHL-1 concentrations were considered too small for reliable analyses (i.e., NIC = 14, MTBI = 17, and STBI = 13). As such, analyses using these three groups were not undertaken. For purposes of exploratory analyses, however, we combined the MTBI and STBI groups (n = 30). In the MTBI and STBI group combined, baseline UCHL-1 concentrations were predictive of symptom change scores on measures of somatic complaints (MLS), neurological complaints (NUC), and anxiety (AXY), accounting for 10.3–16.4% of the variance.

In the NIC group, no meaningful associations were found between baseline biomarker concentrations and symptom change scores on the majority of measures, with the exception of two measures. Baseline tau concentrations were predictive of symptom change scores on measures of anger (ANP; 27.2% variance), and baseline NFL concentrations were predictive of symptom change scores on measures of anxiety (AXY, 13.9% variance).

Discussion

Overall, this study demonstrated that elevated serum tau, NFL, GFAP, and UCHL-1 concentrations within the first 12 months after TBI were associated with the deterioration of a variety of neurobehavioral symptoms that extend into the chronic phase of the recovery trajectory two or more years post-injury. More specifically, in the MTBI group, the strongest effects were found for (a) tau that accounted for 23–26% of the variance in predicting the deterioration of headaches, anxiety, and PTSD re-experiencing symptoms over time, and (b) GFAP that accounted for 23–35% of the variance in predicting the deterioration of depression and somatic complaints.

In the STBI group, the strongest effects were found for (a) tau to predict the deterioration of depression, somatic, and neurological complaints (20–28% variance); (b) NFL to predict depression, cognitive, and neurological complaints (20–32% variance); and (c) GFAP to predict depression, neurological, somatic, and cognitive complaints (20–43% variance).

It is important to appreciate, however, that there were meaningful effects also found for all four serum biomarkers with multiple other neurobehavioral symptoms in the MTBI and STBI groups (e.g., depression, anxiety, post-concussion, and PTSD symptoms; somatic, neurological, and cognitive complaints). While only a handful of these findings were statistically significant (likely because of small sample sizes), these serum biomarkers accounted for 10–19% of the variance in predicting the deterioration of neurobehavioral functioning in these areas and are considered meaningful to understand recovery trajectories.

These findings extend previous research in this area in two important ways. First, these findings demonstrate that serum biomarkers (e.g., tau, NFL, GFAP, UCHL-1) that have been found to be associated with the presence of neurobehavioral dysfunction after TBI (e.g., post-concussion, PTSD, depression, sleep problems, and cognitive impairment) in past cross-sectional studies (e.g., ^{6

–16}) are also associated with the deterioration of neurobehavioral functioning over time. That is, elevated serum biomarkers are associated with the failure of long-term resolution of neurobehavioral symptoms over time, and most alarmingly, a worsening of neurobehavioral symptoms over time.

There is evidence showing that acute activities of these proteins return to baseline levels within two weeks of injury for most individuals; however, for 10–20% of athletes, there are continued elevations of NFL and tau in those with prolonged symptoms.²² Both tau and NFL activity relate to neuronal integrity and function, with NFL and tau being linked to compromised axonal cytoskeleton integrity³² and structural integrity of axons,³³ respectively. In contrast, GFAP, a protein expressed by astrocytes,³⁴ and UCHL-1, which relates to the removal of misfolded and oxidized proteins,³⁵ are elevated acutely, with reductions over the days after a TBI.³⁶ While each of these proteins is involved in slightly different aspects of neuronal functioning, the elevations seen here across all four suggest evidence of longer-term damage, likely related to persistent axonal degeneration and gliosis.

Thus, these findings have notable implications for those individuals who may not demonstrate neurobehavioral symptoms at a clinically meaningful level when first seen by a medical professional after injury, but whose neurobehavioral symptoms may deteriorate to a clinically meaningful level in future. A blood-based panel of these biomarkers may be a useful tool for identifying individuals at risk of poor future outcome, in an effort to provide long-term monitoring of future deterioration, or to provide treatment options as a preventive for deterioration of future neurobehavioral functioning (e.g., resilience training). This is essential, because chronic symptoms after TBIs are linked to neuronal volume loss, which is thought to be a lasting pathology.³⁷

Second, these findings demonstrate that elevated serum tau, NFL, GFAP, and UCHL-1 concentrations within the first 12 months after injury are associated with the deterioration of neurobehavioral symptoms that extend into the chronic phase of recovery two or more years post-injury. In contrast, past longitudinal research has focused primarily on the prognostic utility of serum biomarkers evaluated within the first few hours, days, and/or weeks post-injury to predict outcomes in the first 6–12 months post-injury (e.g., ^18

–21). In a study of patients with mostly severe TBI, tau was observed to remain elevated up to 30 days after injury and to relate to severity of symptoms.³⁸

Our findings suggest that the utility of serum biomarkers to predict future clinical outcomes may extend beyond the acute phase of recovery and may be equally applicable in the subacute and/or chronic phase of recovery, even in those with more mild TBIs. Further, these findings have particular implications for those patients who do not seek medical attention immediately after injury, but rather first seek medical attention many months post-injury. In a study of patients of varying severities who sustained a TBI at least six months prior, NFL at this time point predicted traumatic axonal injury and atrophy.³² These findings suggest that a blood-based panel of these biomarkers may be a useful tool for identifying individuals at risk of poor future outcome not only within the first few days/weeks post-injury, but also within the first 12 months post-injury.

Of further interest, it is important to note that GFAP and UCHL-1 levels were higher in the control group compared with both the MTBI and STBI groups. This finding is inconsistent with previous research that has supported GFAP³⁹ and UCHL-1^{5, 39
–41} as diagnostic markers of TBI. Although it is difficult to explain why the diagnostic utility of these biomarkers is lacking in this sample, our results do demonstrate that GFAP and UCH-L1 may have some prognostic utility after TBI in SMVs.

There are a number of limitations of this study that warrant mention. The sample sizes used in this study were small, which resulted in low statistical power; and a large number of CVs for UCHL-1 were above threshold, which prevented examining the utility of UCHL-1 in the three groups separately. In addition, because of small sample sizes, we were unable to include an injured control group (vs. non-injured control group). Injured control groups are typically used in TBI research to control for the influence of general bodily trauma in an attempt to isolate the influence of brain injury alone. In our TBI sample, 57.9% had been evaluated six or more months post-injury at baseline. While it is unlikely that acute bodily trauma has influenced the findings in this study, the influence of chronic conditions cannot be entirely ruled out.

Further, small sample sizes did not allow us to explore the role of biomarkers in homogenous TBI severity groups (with the exception of the MTBI group). In the STBI group, there was substantial heterogeneity of TBI severity that included complicated MTBI, moderate, severe, and penetrating TBI. In this group, the duration of PTA ranged from <1 min to >30 days (i.e., <1 min = 17.8%; 1–15 min = 7.1%; 16–59 min = 3.6%; 1–24 h = 10.7%; >1–7 days = 21.4%; >7–30 days = 17.9%; >30 days = 10.7%); and the duration of LOC ranged from <1 min to two days (<1 min = 39.3%; 1–15 min = 7.1%; 16–30 min = 3.6%; 31–59 min = 3.6%; 1–24 h = 3.6%; two days = 3.6%; Present Duration Unknown = 39.3%).

In addition, the predominant intracranial abnormalities included diffuse axonal injury = 34.5%; subdural hemorrhage = 20.7%; intraparenchymal hemorrhage = 13.8%; subarachnoid hemorrhage = 1 3.8%; skull fracture/fragments in brain = 10.3%; and gliosis/encephalomalacia = 6.9%.

Nonetheless, study strengths include the use of symptom validity measures to exclude participants for exaggerating symptoms; the use of a well-characterized sample of SMVs that covered the full TBI severity spectrum; the assessment of a wide range of neurobehavioral symptoms; and the use of a longitudinal study design.

Conclusions

This study demonstrated that elevated tau, NFL, GFAP, and UCHL-1 concentrations within the first 12 months of injury are associated with the deterioration of neurobehavioral symptoms that extend into the chronic phase of recovery after TBI. While these results suggest that a blood-based panel of these biomarkers may be a useful tool for identifying individuals at risk of poor future outcome within the first 12 months after injury, future longitudinal studies are needed to replicate these results using larger sample sizes and to link peripheral biomarkers to brain-related changes to better understand possible neuronal pathological processes related to TBIs.

Footnotes

Acknowledgments

The authors would like to express gratitude to the service members and veterans for their time and commitment to participating in this research. The authors would also like to acknowledge the efforts of the larger team of research coordinators, research associates, research assistants, program managers, and senior management who contribute to the DVBIC-TBICoE 15-Year Longitudinal TBI Study.

Disclaimer

The views expressed in this manuscript are those of the authors and do not necessarily represent the official policy or position of the Defense Health Agency, Department of Defense, or any other U.S. government agency.

Authors' Contributions

Rael T. Lange: Conceptualization; Data curation; Formal analysis; Methodology; Supervision; Writing–original draft. Sara Lippa: Conceptualization; Data curation; Supervision; Writing–review and editing. Tracey A. Brickell: Conceptualization; Methodology; Supervision; Writing–review and editing. Jessica Gill: Conceptualization; Data curation; Supervision; Writing–review and editing. Louis M. French: Conceptualization; Supervision; Writing review and editing.

Funding Information

This work was prepared under Contract HT0014-19-C-0004 with DHA Contracting Office (CO-NCR) HT0014 and, therefore, is defined as U.S. Government work under Title 17 U.S.C.§101. Per Title 17 U.S.C.§105, copyright protection is not available for any work of the U.S. Government. For more information, please contact dha.TBICoE.info@mail.mil. UNCLASSIFIED

Author Disclosure Statement

No competing financial interests exist.

References

Shemilt

, Boutin

, Lauzier

, et al. Prognostic value of glial fibrillary acidic protein in patients with moderate and severe traumatic brain injury: a systematic review and meta-analysis. Crit Care Med, 2019; 47(6):e522–ee529; doi: 10.1097/CCM.0000000000003728

Gan

, Stein

, Swanson

, et al. Blood biomarkers for traumatic brain injury: a quantitative assessment of diagnostic and prognostic accuracy. Front Neurol, 2019; 10:446; doi: 10.3389/fneur.2019.00446

Shahim

, Zetterberg

. Neurochemical markers of traumatic brain injury: relevance to acute diagnostics, disease monitoring, and neuropsychiatric outcome prediction. Biol Psychiatry, 2022; 91(5):405–412; doi: 10.1016/j.biolpsych.2021.10.010

Hier

, Obafemi-Ajayi

, Thimgan

, et al. Blood biomarkers for mild traumatic brain injury: a selective review of unresolved issues. Biomark Res, 2021; 9(1):70; doi: 10.1186/s40364-021-00325-5

Wang

, Yang

, Zhu

, et al. An update on diagnostic and prognostic biomarkers for traumatic brain injury. Expert Rev Mol Diagn, 2018; 18(2):165–180; doi: 10.1080/14727159.2018.1428089

Lippa

, Werner

, Miller

, et al. Recent advances in blood-based biomarkers of remote combat-related traumatic brain injury. Curr Neurol Neurosci Rep, 2020; 20(12):54; doi: 10.1007/s11910-020-01076-w

Olivera

, Lejbman

, Jeromin

, et al. Peripheral Total tau in military personnel who sustain traumatic brain injuries during deployment. JAMA Neurol, 2015; 72(10):1109–1116; doi: 10.1001/jamaneurol.2015.1383

Shahim

, Zetterberg

, Tegner

, et al. Serum neurofilament light as a biomarker for mild traumatic brain injury in contact sports. Neurology, 2017; 88(19):1788–1794; doi: 10.1212/WNL.0000000000003912

Pattinson

, Shahim

, Taylor

, et al. Elevated tau in military personnel relates to chronic symptoms following traumatic brain injury. J Head Trauma Rehabil, 2020; 35(1):66–73; doi: 10.1097/HTR.0000000000000485

10.

Pattinson

, Gill

, Lippa

, et al. Concurrent mild traumatic brain injury and posttraumatic stress disorder is associated with elevated tau concentrations in peripheral blood plasma. J Traumatic Stress, 2019; 32(4):546–554; doi: 10.1002/jts.22418

11.

Lippa

, Gill

, Brickell

, et al. Blood biomarkers relate to cognitive performance years after traumatic brain injury in service members and veterans. J Int Neuropsychol Soc, 2021; 27(5):508–514; doi: 10.1017/S1355617720001071

12.

Guedes

, Kenney

, Shahim

, et al. Exosomal neurofilament light: a prognostic biomarker for remote symptoms after mild traumatic brain injury?. Neurology, 2020; 94(23):e2412–e2243; doi: 10.1212/WNL.0000000000009577

13.

Kenney

, Qu

, Lai

, et al. Higher exosomal phosphorylated tau and total tau among veterans with combat-related repetitive chronic mild traumatic brain injury. Brain Inj, 2018; 32(10):1276–1284; doi: 10.1080/02699052.2018.1483530

14.

Gottshall

, Agyemang

, O'Neil

, et al. Sleep quality: a common thread linking depression, post-traumatic stress, and post-concussive symptoms to biomarkers of neurodegeneration following traumatic brain injury. Brain Inj, 2022:36(55):633–643; doi: 10.1080/02699052.2022.2037

15.

Gottshall

, Guedes

, Pucci

, et al. Poor sleep quality is linked to elevated extracellular vesicle-associated inflammatory cytokines in warfighters with chronic mild traumatic brain injuries. Front Pharmacol, 2021; 12:762077; doi: 10.3389/fphar.2021.762077

16.

Motamedi

, Kanefsky

, Matsangas

, et al. Elevated tau and interleukin-6 concentrations in adults with obstructive sleep apnea. Sleep Med, 2018; 43:71–76; doi: 10.1016/j.sleep.2017.1121

17.

Lippa

, Yeh

, Gill

, et al. Plasma tau and amyloid are not reliably related to injury characteristics, neuropsychological performance, or white matter integrity in service members with a history of traumatic brain injury. J Neurotrauma, 2019; 36(14):2190–2199; doi: 10.1089/neu/2018.6269

18.

Hossain

, Mohammadian

, Takala

RSK

, et al. Early levels of glial fibrillary acidic protein and neurofilament light protein in predicting the outcome of mild traumatic brain injury. J Neurotrauma, 2019; 36(10):1551–1560; doi: 10.1089/neu.2018.5952

19.

Thelin

, Al Nimer

, Frostell

, et al. A serum protein biomarker panel improves outcome prediction in human traumatic brain injury. J Neurotrauma, 2019; 36(20):2850–2862; doi: 10.1089/neu.2019.6375

20.

Shahim

, Gren

, Liman

, et al. Serum neurofilament light protein predicts clinical outcome in traumatic brain injury. Sci Rep, 2016; 6:36791; doi: org/10.1038/srep36791

21.

Shahim

, Tegner

, Marklund

, et al. Neurofilament light and tau as blood biomarkers for sports-related concussion. Neurology, 2018; 90(20):e1780–e1788; doi: 10.1212/WNL.0000000000005518

22.

Pattinson

, Meier

, Guedes

, et al. Plasma biomarker concentrations associated with return to sport following sport-related concussion in collegiate athletes-a Concussion Assessment, Research, and Education (CARE) Consortium Study. JAMA Netw Open, 2020; 3(8):e2013191; doi: 10.1001/jamanetworkopen.2020.13191

23.

Corrigan

, Bogner

. Initial reliability and validity of the Ohio State University TBI Identification Method. J Head Trauma Rehabil, 2007; 22(6):318–329; doi: 10.1097/01.HTR.0000300227.67748.77

24.

Ben-Porath

, Tellegen

MMPI-2-RF (Minnesota Multiphasic Personality Inventory-2 Restructured Form): manual for administration, scoring, and interpretation. University of Minnesota Press: Minneapolis, MN; 2011.

25.

Cicerone

, Kalmar

. Persistent postconcussion syndrome: The structure of subjective complaints after mild traumatic brain injury. J Head Trauma Rehabil, 1995; 10(3):1–7; doi: 10.1097/0000119906000-00002

26.

Vanderploeg

, Silva

, Soble

, et al. The structure of postconcussion symptoms on the Neurobehavioral Symptom Inventory: a comparison of alternative models. J Head Trauma Rehabil 201;30(1):1-11; doi: 10.1097/HTR.0000000000000009

27.

Blanchard

, Jones-Alexander

, Buckley

, et al. Psychometric properties of the PTSD checklist (PCL). Behav Res Ther, 1996; 34(8):669–673; doi: 10.1016/0005-7967(96)00033-2

28.

American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. American Psychiatric Association: Washington, DC; 1994.

29.

Belanger

, Lange

, Bailie

, et al. [Formula: see text] Interpreting change on the neurobehavioral symptom inventory and the PTSD checklist in military personnel. Clin Neuropsychol, 2016; 30(7):1063–1073; doi: 10.1080/13854046.2016.1193632

30.

Corrigan

, Bogner

, Mellick

, et al. Prior history of traumatic brain injury among persons in the Traumatic Brain Injury Model Systems National Database. Arch Phys Med Rehabil, 2013; 94(10):1940–1950; doi: 10.1016/j.apmr.2013.05.018

31.

Trotta

, Ekanayake

, Ettenhofer

, et al. Intracranial abnormalities are associated with fewer self-reported symptoms in military service members following moderate-to-severe traumatic brain injury. J Head Trauma Rehabil, 2021; 36(3):164–174; doi: 10.1097/HTR.0000000000000637

32.

Shahim

, Politis

, van der Merwe

, et al. Time course and diagnostic utility of NfL, tau, GFAP, and UCH-L1 in subacute and chronic TBI. Neurology, 2020; 95(6):e623–e636; doi: 10.1212/0000000000009985

33.

Castellani

, Perry

. Tau biology, tauopathy, traumatic brain injury, and diagnostic challenges. J Alzheimers Dis. 2019; 67(2):447–467; doi: 10.3233/JAD-180721

34.

McMahon

, Panczykowski

, Yue

, et al. Measurement of the glial fibrillary acidic protein and its breakdown products GFAP-BDP biomarker for the detection of traumatic brain injury compared to computed tomography and magnetic resonance imaging. J Neurotrauma, 2015; 32(8):527–533; doi: 10.1089/neu.2014.3635

35.

Gong

, Leznik

. The role of ubiquitin C-terminal hydrolase L1 in neurodegenerative disorders. Drug News Perspect, 2007; 20(6):365–370; doi: 10.1358/dnp.2007.20.6.1138160

36.

Giza

, McCrea

, Huber

, et al. Assessment of blood biomarker profile after acute concussion during combative training among US military cadets: a prospective study from the NCAA and US Department of Defense CARE Consortium. JAMA Netw Open, 2021; 4(2):e2037731; doi: 10.1001/jamanetworkopen.2020.37731

37.

Zhang

, Puvenna

, Janigro

Biomarkers of traumatic brain injury and their relationship to pathology. In: Translational Research in Traumatic Brain Injury. ( Laskowitz

, Grant

eds.) CRC Press/Taylor and Francis Group: Boca Raton, FL; 2016

38.

Bogoslovsky

, Gill

, Jeromin

, et al. Fluid biomarkers of traumatic brain injury and intended context of use. Diagnostics (Basel). 2016; 6(4); doi: 10.3390/diagnosics6040037

39.

FDA Decision Memorandum. Evaluation of Automatic Class III Designation for Benyon Brain Trauma Indicator; 2018

40.

Papa

, Lewis

, Silvestri

, et al. Serum levels of ubiquitin C-terminal hydrolase distinguish mild traumatic brain injury from trauma controls and are elevated in mild and moderate traumatic brain injury patients with intracranial lesions and neurosurgical intervention. J Trauma Acute Care Surg, 2012; 72(5):1335–1344; doi: 10.1097/TA.0b013e3182491e3d

41.

Diaz-Arrastia

, Wang

, Papa

, et al. Acute biomarkers of traumatic brain injury: relationship between plasma levels of ubiquitin C-terminal hydrolase-L1 and glial fibrillary acidic protein. J Neurotrauma, 2014; 31(1):19–25; doi: 10.1089/neu.2013.3040